Bond Math Tells One Analyst That Ultra-Long Treasuries Would Save the U.S. MoneyBy
Weighing 50-year bond versus a 30-year rolled into a 20-year
Average maturity of U.S. debt remains under six years
There are plenty of reasons why the U.S. Treasury may eventually decide against introducing an ultra-long bond. But cost shouldn’t be one of them, according to Niso Abuaf, a Wall Street financial strategist since the mid-1980s.
A basic exercise in bond math shows that the government would likely save money by stretching its yield curve to the 50-year mark, said Abuaf, who heads the financial strategy group at Ramirez & Co. in New York. The analysis involves comparing the theoretical cost of issuing a 50-year bond today with the expense of alternatives such as issuing a 30-year bond and then rolling that into a 20-year obligation.
The approach requires an interpolated 50-year Treasury yield and the zero-coupon yield curve. It determines the net present value of a 50-year Treasury and the coupon rate for a 20-year bond issued 30 years from now that would give the 30-year/20-year combination the same net present value.
“We call that the breakeven rate,” said Abuaf, an economist who held similar roles at Salomon Brothers and Credit Suisse. “If the 20-year rate 30 years from now is going to be greater than that breakeven, then issuing 50-years today would be cheaper, and conversely.”
Abuaf’s analysis assumes that a 50-year Treasury sold today would yield about 3.4 percent, based on yields of bonds issued by the Tennessee Valley Authority and Microsoft Corp. The resulting breakeven rate for a 20-year bond sold to replace a maturing 30-year is 4.88 percent.
At that rate, the issuer would be indifferent between selling 50-year bonds today and the alternative combination. The next step is to determine the chance that yields exceed the breakeven rate.
This part “is not math anymore,” Abuaf said. “This is: What does your macroeconomic intuition say and whether you think that breakeven will be breached or not.”
Assuming economic growth of 2 percent -- the post-crisis average -- or the long-term historical average of about 3 percent, and inflation around 2 percent, then 4 percent to 5 percent “plus a reasonable term premium” is the long-term forecast for yields on Treasuries, he said. In that scenario, “my intuition says it’s a good policy thing to do,” Abuaf said.
Treasury Secretary Steven Mnuchin has said ultra-long issuance makes sense, and primary dealers were asked for their views on the prospect in connection with this month’s refunding. Dealers have been cool to the idea, with the Treasury Borrowing Advisory Committee saying evidence of “strong or sustainable demand” is lacking past the 30-year mark.
Abuaf’s analysis rests to a large degree on the 50-year yield assumption, which he concedes is questionable because of uncertainty about the size and frequency of any ultra-long program Treasury might introduce.
There are other reasons for the U.S. to consider creating an ultra-long curve point, including that “less creditworthy countries” like Mexico have done it, he said.
And although the average maturity of Treasury debt at almost six years is near the high end of its range since the 1980s, “hopefully we will have a life as a nation that is far longer than that.”